gym_multiagent

Dependency

OpenAI Gym

Installation

git clone https://github.com/kbys-t/gym_MA.git
cd gym_MA
pip install -e .

How to use

1. First of all, import gym_multiagent

2. Select environment from ["MoveFormMA-v0", "MoveFormMA-v1"]

ENV_NAME = "MoveFormMA-v0"
env = gym.make(ENV_NAME)

1. Prepare agents

age = []
for i in range(env.AGE_NUM):
    age.append( some_agent() )

1. Send listed actions and Get listed observations and rewards for all agents together

observation, reward, done, info = env.step(action)
for i in range(env.AGE_NUM):
    age[i].set_State(observation[i])
    age[i].update(reward[i])
    action[i] = age[i].get_Action()